“…For the above-mentioned problems, at present, domestic and international researches mainly focused on a specific environment or on specific target recognition, specific scene understanding under different conditions (Ilbeygi and Shah-Hosseini, 2012;Novak et al, 1997;Liu et al, 2011;Kulkarni et al, 2013;Lim et al, 2014;Huang et al, 2014), but have seldom carried out feature extraction, target recognition, and semantic understanding of scene for being fragmented, contaminated grayscale images and other fuzzy images. To achieve the accurate semantic understanding for the scene of fuzzy image, it must implement the accurate target recognition to the image scene.…”